First gen ColorHug broken?

[Sven Arvidsson]

Hi,

My first generation ColorHug seems to have developed quite a problem:

On my Debian systems it is not detected in GNOME settings or by
gcm-picker. It is however detected by colorhug-flash and offers to
upgrade the firmware from 1.2.0. But the upgrade process fails with
these errors:

failed to contact ColorHug
usb:00:0c: Invalid read: retval=0x0b [Unknown command for bootloader]
cmd=0x0b [get-serial-number](expected 0xb [get-serial-number]) len=2
(expected 6 or 64)

Failed to write the firmware
usb:00:0c: USB error on device 04d8:f8da: No such device (it may have
been disconnected)

A Fedora 24 Live USB detects the ColorHug, but fails to
calibrate. Either GNOME settings crashes, colord crashes or
calibration ends with "An internal error occurred that could not be
recovered".

gcm-picker (On Fedora 24) reports "failed to get sample".

I have tried different computers, different USB ports, and different
cables without success.

Broken hardware, anything else I could try?


--
Cheers,
Sven Arvidsson
http://www.whiz.se

[Peter D'Hoye]

failed to contact ColorHug
usb:00:0c: Invalid read: retval=0x0b [Unknown command for bootloader]
cmd=0x0b [get-serial-number](expected 0xb [get-serial-number]) len=2
(expected 6 or 64)

Failed to write the firmware
usb:00:0c: USB error on device 04d8:f8da: No such device (it may have
been disconnected)

Had the same thing... download the firmware from the site and flash manually, that worked for me

Peter

[Sven Arvidsson]

On Sat, 2016-10-22 at 19:16 +0200, Peter D'Hoye wrote:
> Had the same thing... download the firmware from the site and flash
> manually, that worked for me

No luck here. I get the same error message when downloading the
firmware manually.

[Richard Hughes]

On 22 October 2016 at 16:41, Sven Arvidsson <s...@whiz.se> wrote:
> usb:00:0c: Invalid read: retval=0x0b [Unknown command for bootloader]
> cmd=0x0b [get-serial-number](expected 0xb [get-serial-number]) len=2
> (expected 6 or 64)

It looks stuck in bootloader mode.

> Failed to write the firmware
> usb:00:0c: USB error on device 04d8:f8da: No such device (it may have
> been disconnected)

Can you try a different port/cable/computer for flashing back to firmware mode?

Richard.

[Sven Arvidsson]

On Mon, 2016-11-21 at 11:08 +0000, Richard Hughes wrote:
> It looks stuck in bootloader mode.
>
> > Failed to write the firmware
> > usb:00:0c: USB error on device 04d8:f8da: No such device (it may
> > have
> > been disconnected)
>
> Can you try a different port/cable/computer for flashing back to
> firmware mode?

Hi,

I have tried three different systems now, and two cables, The "no such
device" error is gone. but I keep getting

usb:00:0c: Invalid read: retval=0x0b [Unknown command for bootloader]
cmd=0x0b [get-serial-number](expected 0xb [get-serial-number]) len=2
(expected 6 or 64)

Resetting the device with boot-flash and set-flash-success after failed
flash works, but the error persists if I try again.

[Richard Hughes]

On 28 November 2016 at 13:29, Sven Arvidsson <s...@whiz.se> wrote:
> Resetting the device with boot-flash and set-flash-success after failed
> flash works, but the error persists if I try again.

I'm a bit confused now. Could you try updating with fwupmgr and see if
that's able to set the flash success bits correctly. Thanks.

Richard.

[Sven Arvidsson]

On Mon, 2016-11-28 at 14:37 +0000, Richard Hughes wrote:
> I'm a bit confused now. Could you try updating with fwupmgr and see
> if
> that's able to set the flash success bits correctly. Thanks.

fwupmgr does work and updates the firmware correctly, the ColorHug is
now detected in gnome-control-center.

But that's where my luck runs out and I'm seeing the same problems as a
I described with the Fedora live USB:

gnome-control-center gives an error when calibration is about to start:

(gnome-control-center:14054): color-cc-panel-WARNING **: failed to
start calibrate:
GDBus.Error:org.freedesktop.ColorHelper.FailedToFindSensor: failed to
lock: usb:00:02: Invalid read: retval=0x0a [Invalid value] cmd=0x35
[get-remote-hash] (expected 0x35 [get-remote-hash]) len=64 (expected 22
or 64)

and if I try again g-c-c segfaults:

(gnome-control-center:14054): color-cc-panel-CRITICAL **:
cc_color_calibrate_setup: assertion 'calibrate->priv->device_kind !=
CD_SENSOR_CAP_UNKNOWN' failed

Trying dispcal I get this error:

19:14:26 VTE: error: new_disprd failed with 'Instrument Access
Failed'

Tried different systems, different ports, different cables.

[Graeme Gill]

Sven Arvidsson wrote:

> Trying dispcal I get this error:
>
> 19:14:26 VTE: error: new_disprd failed with 'Instrument Access
> Failed'

To get more detail there, you would need to run something like
ArgyllCMS's "spotread -D8"

Graeme Gill.

[Cheng-Chia Tseng]

I have met similar situations as well.

The color management from GNOME control center always failed on f24 and even f25.

However, DispCAL worked quite well. Maybe you can try it instead.

Graeme Gill <gra...@argyllcms.com>於 2016年11月30日 週三，上午7:29寫道：

--
You received this message because you are subscribed to the Google Groups "colorhug-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to colorhug-user...@googlegroups.com.
Visit this group at https://groups.google.com/group/colorhug-users.
For more options, visit https://groups.google.com/d/optout.

[Richard Hughes]

On 29 November 2016 at 13:52, Sven Arvidsson <s...@whiz.se> wrote:
> lock: usb:00:02: Invalid read: retval=0x0a [Invalid value] cmd=0x35
> [get-remote-hash] (expected 0x35 [get-remote-hash]) len=64 (expected 22
> or 64)

I've pushed a new firmware just now that disables the
return-error-on-unset functionality. If you install the file from
https://secure-lvfs.rhcloud.com/lvfs/device/40338ceb-b966-4eae-adae-9c32edfcc484
this error should disappear. Just use "fwupdmgr install
hughski-colorhug-1.2.6.cab" or "fwupdmgr refresh && fwupdmgr update"
to update the firmware on any recentish Linux. If you need to install
this using colorhug-cmd, just explode the .cab file and write the
binary file.

Richard.

[Sven Arvidsson]

On Fri, 2016-12-02 at 13:19 +0000, Richard Hughes wrote:
> I've pushed a new firmware just now that disables the
> return-error-on-unset functionality. If you install the file from
> https://secure-lvfs.rhcloud.com/lvfs/device/40338ceb-b966-4eae-adae-9
> c32edfcc484
> this error should disappear. Just use "fwupdmgr install
> hughski-colorhug-1.2.6.cab" or "fwupdmgr refresh && fwupdmgr update"
> to update the firmware on any recentish Linux. If you need to install
> this using colorhug-cmd, just explode the .cab file and write the
> binary file.

Awesome! 

Calibration now works using Low and Medium settings, if I try high,
calibration fails with this error:

gnome-control-center:26224): color-cc-panel-WARNING **: calibration
failed with code 5: Gamma correction table was non-monotonic:
0.000000,0.000000,0.000000 0.181250,0.090000,0.016875
0.168750,0.080000,0.002500 0.209375,0.138750,0.041250
0.234375,0.160000,0.067500 0.278125,0.196875,0.100000
0.318750,0.240000,0.138750 0.362500,0.280000,0.177500
0.412500,0.320000,0.222500 0.462500,0.360000,0.270625
0.521875,0.400000,0.312500 0.568750,0.440000,0.341875
0.621875,0.480000,0.390000 0.671875,0.520000,0.431875
0.728125,0.560000,0.483125 0.778125,0.600000,0.528125
0.831250,0.640000,0.582500 0.881250,0.676875,0.630625
0.937500,0.726250,0.697500 0.962500,0.760000,0.770625
1.000000,0.800000,0.837500

I'm still hitting the assert on any subsequent calibrations, but I see
there's already an open bug for that.